Abstract
Abstract: :
Purpose: Interference between reflections from the front and back of the tear film causes oscillations in the reflectance spectrum, whose ‘frequency’ is proportional to tear film thickness (King–Smith et al., 2000, IOVS 41, 3348). The purpose of this study was to investigate whether spectral oscillations of different frequency can be detected, corresponding to reflections from different depths behind the air surface. For example, reflections from the aqueous–mucus interface and from the epithelial surface, could generate spectral oscillations of different frequencies, which could combine to generate 'beats' in the reflectance spectrum. Methods: The analysis used the 200 spectra (435 to 1053 nm) with the best signal/noise ratio recorded from 30 women (mean age 47). Measurement spot size was nominally 28 x 17 um and exposure of 0.5 s was taken about 2 s after a blink. A program was written to make a least squares fit to the oscillations based on one or two frequency components. These components could be ‘undamped’ (constant amplitude as a function of wave number = 1/wavelength, corresponding to a smooth, sharp surface) or ‘damped’ (amplitude decays exponentially as a function of wave number, corresponding to a rough or diffuse surface). Three parameters (amplitude, frequency and phase) were optimized for an undamped component, whereas a fourth parameter (decay rate) was needed for the damped fit. Results: Root–mean–square deviation between measured and fitted reflectance spectra for different fits are: 1 damped component (4 parameters), 0.296%; 2 undamped components (6 parameters) 0.265%; 1 damped and 1 undamped components (7 parameters) 0.116%; 2 damped components (8 parameters) 0.112%. Thus 7 parameters gives a considerably better fit than fewer parameters, but adding another parameter causes little further improvement. For 193 of the 200 spectra, the damped component corresponds to a shallower surface than the undamped component (median thickness difference 0.26 um). The median amplitude, (at 800 nm), of the damped component is 295% of the undamped component. Conclusions: The reflection from the aqueous–mucus surface is expected to be relatively weak (because of the low concentration of mucins in mucus) and the corneal surface is expected to give a rough reflection. Thus the interpretation that the strong–shallow–rough (or diffuse) and weak–deep–smooth components come, respectively, from the aqueous–mucus boundary and corneal surface, seems unlikely. Other interpretations based on reflections from different parts of the corneal surface (e.g., tips and bases of microplicae) will be discussed.
Keywords: cornea: tears/tear film/dry eye • cornea: surface mucins • optical properties